Self-adaptive and self-aware mobile-cloud hybrid robotics

Many benefits of cloud computing are now well established, as both enterprise and mobile IT has been transformed by cloud computing. Backed by the virtually unbounded resources of cloud computing, battery-powered mobile robotics can also benefit from cloud computing, meeting the demands of even the most computationally and resource-intensive tasks. However, many existing mobile-cloud hybrid tasks are inefficient in terms of achieving objectives like minimizing battery power consumption and network bandwidth usage, which form a tradeoff. To counter this problem we propose a technique based on offline profiling, that allows class, method and hybrid level configurations to be applied to MC hybrid robotic tasks and measures, at runtime, how well the tasks meet these two objectives. The optimal configurations obtained from offline profiling are employed to make decisions at runtime. The decisions are based on: 1) changing the environment (i.e. WiFi signal level variation), and 2) itself in a changing environment (i.e. actual observed packet loss in the network). Our experimental evaluation considers a Python-based foraging task performed by a battery-powered and Raspberry Pi controlled Thymio robot. Analysis of our results shows that self-adaptive and self-aware systems can both achieve better optimization in a changing environment (signal level variation) than using static offloading or running the task only on a mobile device. However, a self-adaptive system struggles to perform well when the change in the environment happens within the system (network congestion). In such a case, a self-aware system can outperform, in terms of minimizing the two objectives.